We undertook the first case control study of histologically confirmed esophageal candidiasis (EC).
A computer search from July 2012 through February 2015 identified 1,011 esophageal specimens, including 40 cases of EC and 20 controls.
The EC incidence was 5.2%; it was associated with immunosuppression and endoscopic white plaques and breaks. Smoking was a predisposing factor, and alcohol was protective. EC had no unique symptoms, and 54% of endoscopic reports did not suspect EC. Important histologic clues included superficial and detached fragments of desquamated and hyper-pink parakeratosis, acute inflammation, intraepithelial lymphocytosis, dead keratinocytes, and bacterial overgrowth. Thirty percent had no neutrophilic infiltrate. Pseudohyphae were seen on H&E in 92.5% (n = 37/40). "Upfront" periodic acid-Schiff with diastase (PAS/D) on all esophageal specimens would have generated $68,333.49 in patient charges. Our targeted PAS/D strategy resulted in $13,044.87 in patient charges (cost saving = 80.9%, $55,288.62).
We describe the typical morphology of EC and recommend limiting PAS/D to cases where the organisms are not readily identifiable on H&E and with at least one of the following: (1) ulcer, (2) suspicious morphology, and/or (3) clinical impression of EC.